1 Mars MOURA magnetometer demonstration for high resolution mapping on 2 terrestrial analogues 3 4 Marina Diaz-Michelena1, Rolf Kilian2,3, Ruy Sanz1, Francisco Rios2, Oscar Baeza2 5 6 1 Payloads and Space Sciences Department, INTA, Ctra. Torrejón – Ajalvir km 4, 28850 Torrejón de 7 Ardoz, Spain 8 2 Geology Department, University of Trier, Behringstrasse, 54286 Trier, Germany 9 3 University of Magellanes, Punta Arenas, 10 11 Abstract 12 Satellite-based magnetic measurements of Mars indicate complex and very strong 13 magnetic anomalies, which led to an intensive and long-lasting discussion about their 14 possible origin. To make some progress in the investigation of the origin of these 15 anomalies MOURA vector magnetometer was developed for in situ measurements on 16 Mars. In this work we propose the utilisation of such instrument for future planetary 17 on ground surveys. The proof of its suitability is done by testing on various terrestrial 18 analogues characterised by most distinct magnetic anomalies of their basement rocks: 19 (1) A magnetite body of EL Laco (up to + 110,000 nT) and its transition to surrounding 20 (< + 2,000 nT) in the Northern Andes of Chile showing the highest local 21 magnetic anomalies. The magnetite-bearing ore body has highly variable local 22 anomalies due to their complex formation history where a significant dispersion in 23 paleo-orientations has been previously reported, while our vector data show relatively 24 uniform and probably induced declinations. (2) A basaltic spatter cone of the Pali Aike 25 volcanic field, in Southern Chile, was characterised by very strong magnetic anomalies 26 along the crater rim (up to + 12,000 nT), controlled by the amount of single domain 27 magnetites in the ground mass of the . Due to their strong remanent signature 28 paleodeclinations of the and reorientations of collapsed blocks could be 29 constrained by the vector data. (3) The Monturaqui meteorite crater (350 m diameter), 30 in Northern Chile, shows significant variations of its anomalies (from - 2,000 to > + 31 6,000 nT) in restricted areas of several square metres along its crater rim related to 32 unexposed iron-bearing fragments of the impactor while its granitic and ignimbritic 33 target rocks exhibit only very weak anomalies. (4) An area with several amphibolitic 34 dykes which cross-cut a Cretaceous granitoid in the southernmost , where a 35 decimetre-scale mapping was performed. In this case, pyrrhotite is the only magnetic 36 carrier. It was formed during hydrothermal processes within the dykes. Very low (+ 40 37 to + 120 nT) positive magnetic anomalies clearly depict the amount of 1 to 4 vol.% 38 pyrrhotite in these dykes, which is important as a mineralogical indicator as well as to 39 detect associated gold and copper enrichment. 40 41 (355 words) 42 43 Keywords: Andes, El Laco, Monturaqui, Pali Aike, Patagonian Batholith, continental 44 crust, martian crust, granite, , dykes, volcanic crater, magnetic anomalies, 45 magnetic carrier, magnetic susceptibility, ground magnetic survey, caesium 46 magnetometer, MOURA magnetometer, Koenigsberger ratios, plutonic rocks, impact 47 crater, iron meteorite, pyrrhotite, magnetite, hydrothermal mineralization, gold, 48 copper.

1

1 1. Introduction 2 3 Mars magnetic field has been exhaustively measured between 100 and 440 km 4 altitude by Mars Global Surveyor (Acuña et al., 1998; Connerney et al., 2005; 5 Morschhauser et al., 2014). These data show that the magnetic anomalies of the 6 martian crust are up to 20 times higher than those of the Earth (Scott and Fuller, 7 2004). However, a profound understanding of the magnetic signature of the martian 8 crust would require mapping at different altitudes and consequently with a different 9 magnetic zoom apart from further petrological analyses. Since Mars presents a very 10 low dense atmosphere, aeromagnetic surveys are not achievable in the short term. 11 Thus, on ground magnetometry with landers and rovers seems to be the most 12 immediate feasible technology to complement the satellite measurements. MOURA 13 magnetometer was developed by INTA in the context of MetNet Precursor Mission to 14 perform vector magnetometry and gradiometry during on ground prior to rover-based 15 surveys on extra-terrestrial planets, like Mars. The instrument has a very low mass (72 16 g), a scalable range, high precision, low detectable fields and noise, and is capable to 17 work in the very hard environmental conditions of Mars. Diaz-Michelena et al. (2015 a) 18 describes MOURA main technical details, the calibration, as well as the demonstration 19 to measure the absolute magnetic field and its temporal variations. 20 The first objective of the present work is to demonstrate the capability of the 21 miniaturized MOURA instrument in a real context of terrestrial analogues surveys by 22 means of the inter-comparison with the data of a scalar caesium reference 23 magnetometer (Diaz-Michelena and Kilian 2013). To do this, four different sites with a 24 wide variability in the intensity of their magnetic signatures have been selected. A 25 second objective is the magnetic investigation of these sites and their implication as 26 terrestrial analogues of Mars. A further objective is the potential of high resolution 27 mapping to show the lowest magnetic contrasts in the terrain and their correlation 28 with the distinct magnetic carriers responsible for their signatures (Acuña et al., 1998; 29 Connerney et al., 2005; Lillis et al., 2013). 30 Our selected sites include magnetic anomalies from complex geological 31 environments where e.g. noise factors, geometrical characteristics of non-exposed 32 rock units and effects of the terrain relief can obscure partly the interpretations 33 concerning the kind and magnetic effects of non-exposed rock units. Thus a modelling 34 of the magnetic anomalies is desirable in general to improve the interpretation of 35 geometrical and compositional effects of non-exposed rocks (e.g. Eppelbaum et al., 36 2015; Eppelbaum and Mishne, 2011; Ialongo et al., 2014). Since we are aware of this 37 problematics, primarily we focus on the interpretation of the effects of distinct types 38 and compositions of exposed rocks concerning the observed magnetic anomalies. 39 Future more detailed studies of the investigated sites should include the above 40 mentioned magnetic modelling which is out of the scope of this magnetometer 41 demonstration. 42 43 44 2. Methodology 45 46 2.1. Magnetic Instrumentation 47

2

1 Two different magnetometers have been used for the present surveys: a conventional 2 caesium scalar magnetometer: model G-858 MagMapper by Geometrics and MOURA 3 vector magnetometer designed and developed by INTA MetNet team for Mars 4 exploration. 5 G-858 is taken as the reference magnetometer because it is a well-established 6 hand-held instrument (8 - 9 kg) for magnetic surveys (ordnance, archaeology, 7 environmental, mineralogy and petroleum prospection). It has 8 hours autonomy, 8 provides a suitable contrast related to magnetic anomalies (8 pT/√Hz) and good 9 stability covering the range of the Earth magnetic field with a dynamic range between 10 20,000 to 100,000 nT. It also has several modes of operation: continuous and discrete 11 to allow users to plan the prospections grids. In contrast to MOURA, this 12 magnetometer only provides the intensity of the total magnetic field, its performance 13 is dependent on the orientation of the head respect to the field, which changes 14 significantly in the latitude range of the present survey, and it is restricted to areas 15 with gradients higher than 20,000 nT/m (Table 1). 16 MOURA is a vector magnetometer with two 3-axes magnetic sensors of 17 Anisotropic MagnetoResistance (AMR) by Honeywell to build up a compact and 18 miniaturized instrument (72 g mass and 67.5 cm3) for Mars exploration. The power 19 consumption is limited to 400 - 430 mW, so it can operate during more than 10 hours 20 with commercial batteries, and insignificant increase of weight to the user (3 x 25 g). 21 The instrument is also designed for continuous or discrete modes of operation, it can 22 work in every orientation with no incidence in the performance, and it is practically 23 immune to gradients due to the small size of the transducer (µm-size). 24 The characteristics are designed for Mars surface environment (- 90 to + 20 ⁰C in 25 operation, - 120 to + 125 ⁰C in storage, and a total irradiance dose of 15 krad/s). The 26 resolution is limited by the transducer (0.2 nT) and the range is adapted to that of the 27 Earth geomagnetic field ± 65,000 nT with an extended range in the auto-offset 28 compensation mode of ± 130,000 nT (Table 1, Diaz-Michelena et al., 2015 a). 29 30 2.2 Track performance 31 32 The tracks have been defined to cover most of the relevant geological features of the 33 selected areas. Continuous and discrete modes have been selected depending on the 34 characteristics and heterogeneity of the sites. For example, the continuous mode has 35 been applied in extended areas in order to have more flexibility and speed to move. In 36 areas with very small-scale heterogeneities the discrete mode has been preferred. In 37 these cases between 5 and 7 measurements have been taken and averaged per point. 38 An advantage of this mode is that it can be measured directly on ground or at a fixed 39 distance above ground. 40 The positions of measuring points and tracks have been georeferenced by a 41 Garmin 62s GPS. The GPS tracks have been used to derive the orientation. For some 42 relatively small mapping areas, like Bahía Glaciares (Site 4; Fig. 1), a grid of 20 x 20 m 43 was previously defined by a tape measure since the error of GPS positions could be in 44 the order of several metres. In these cases, the lines have been used for the 45 orientation. 46 It has to be taken into account that the different data have been obtained from 3

1 multiple instruments individually without an automatic synchronism. Therefore, all the 2 acquisition units have been manually synchronised, and the sequence of measurement 3 has been done systematically as follows: 4 1) Marking selected measurement point. 5 2) GPS measurement with time stamp (with > 6 satellites at direct sight). 6 3) Removal of the GPS from measurement point to avoid magnetic contamination by 7 this device. 8 4) Magnetic measurement with time stamp (for G-858). 9 5) Magnetometer (three axes), accelerometer (three axes), temperature 10 measurement with time stamp (for MOURA). 11 The data files have been pre-processed manually (preliminary corrections of the 12 GPS data with the above-described information). Ad-hoc software has been performed 13 to include the temperature and tilt angles correction of MOURA data, to subtract the 14 Earth geomagnetic field with respect to total intensities in the case of G-858 and each 15 vector components in the case of MOURA, and to plot the different magnitudes of the 16 processed data. 17 Local magnetic field anomalies have been calculated with respect to the 18 International Geomagnetic Reference Field (IGRF) averaged for the month of the 19 surveys. At single sites the surveys were performed during less than 2 hours for which 20 global magnetic field data from a base station and/or from the next magnetic 21 observatories with minute-resolution have been considered for reference (Argentine 22 Islands near Antarctic Peninsula, Port Stanley and Easter Island and Huancayo ~ 1400 23 km). However, since temporal variations during all survey intervals of single sites were 24 less than ± 10 nT (quiet days), site specific data corrections have not been applied. 25 In all the surveys it has been systematically calculated the correlation parameters 26 between the scalar data of the two magnetometers (G-858 an MOURA) as well as 27 between the vector data of the two separate sensors of MOURA. 28 29 30 2.3 Selected sites: mineralogical and geological context 31 The selected test sites for the on ground survey are situated in or near the Southern 32 Andes between latitudes 20°S to 52°S (Fig. 1). 33 Required general site characteristics are that a) exposed rocks are relatively 34 unaltered, b) high resolution grids with scales of metres to centimetres can be 35 performed, and c) exposed rocks are representative for a large number of martian 36 surface rocks. 37 SITE 1 “El Laco”. Four large magnetite bodies with a total estimated ore resource of 38 500 million tons crop out around El Laco in the Central Andes (Fig. 2A; Alva- 39 Valdivia et al., 2003; Naranjo et al., 2010). Together with the iron ore deposits of 40 Kiruna (e.g. Jonnsson et al., 2013) they represent worldwide unique examples for very 41 strong local magnetic anomalies, which may be comparable to that observed in some 42 areas of the southern Noachian highlands of Mars (Connerney et al., 2005; Lillis et al., 43 2013). 44 The selected area with an extension of 0.2 x 0.4 km is situated at the northern 45 margin of the El Laco Sur outcrop (23°50’17’’S; 67°29’27’’ W; 4720 m elevation; Figs. 1

4

1 and 2), at the transition between magnetite-bearing ores and early Pleistocene 2 andesitic flows which are partly covered by pyroclastic deposits with up to 5 m 3 thickness. The mapping area has not been modified by iron ore mining. 4 Sernageomin (Servicio Nacional de Geología y Minería of Chile; Naranjo et al., 5 2010) performed aeromagnetic surveys and constructed maps of the anomaly. They 6 reflect a dipolar anomaly according to the isodynamic lines of the map with intensities 7 in the order of ± 150 nT (Alva-Valdivia et al., 2003). However, there is no information 8 available concerning the tracks of these aeromagnetic surveys: spacing and altitudes 9 above ground, because of which, the data are only considered for completeness. 10 Besides, on ground magnetic surveys have not been done previously. 11 Fission-track dating of apatite grown within the magnetites gave an age of 2.1 ± 12 0.1 Ma (Maksaev et al., 1988). A whole rock age of the from the host rocks of 13 Pico Laco is 2.0 ± 0.3 Ma (Gardeweg and Ramírez, 1985). 14 The origin of the magnetite bodies has been strongly debated. A combined 15 magmatic and hydrothermal origin was proposed by Alva-Valdivia et al. (2003), Sillitoe 16 and Burrows (2002), and Velasco and Tornos (2012). This is based on the fact that field 17 and petrograhic evidences suggest that some magnetites have a primarily magmatic 18 texture, whereas others show features which indicate a formation during a 19 hydrothermal triggered re-emplacement of andesitic lava flows. Trace element 20 compositions of the latter magnetite type are not compatible with a magmatic origin. 21 For example, Dare et al. (2014) document that these magnetites are characterized by 22 high Ni/Cr ratios, depleted in Ti, Al, Cr, Zr, Hf and Sc, and show an oscillating zoning of 23 Si, Ca, Mg and rare earth elements. In contrast, oxygen isotope data (δ18O of +2 to +4) 24 of many magnetites support a magmatic rather than hydrothermal origin (Jonnsson et 25 al., 2013). 26 Microscopy studies under reflected light as well as temperature dependent 27 susceptibility measurements and isothermal remanent magnetization (IRM) acquisition 28 show that low Ti-magnetite and/or maghemite are the magnetic carriers (Alva-Valdivia 29 et al., 2003). Sometimes ilmenite-hematite minerals appear in significant amounts. 30 Grain sizes range from a few microns up to several millimetres. Hysteresis 31 measurements of Alva-Valdivia et al. (2003) of seven ore samples from El Laco Sur 32 point to pseudo-single-domain status and show a large range of Koenigsberger ratios 33 (Q-ratios from 0.02 to > 1000). 34 Paleomagnetic data show distinct local declinations indicating a complex 35 crystallization history, probably during different geomagnetic field orientations (Alva- 36 Valdivia et al., 2003). 37 SITE 2 “Pali Aike”. Lava sheets with volcanic spatter cones represent a common feature 38 in many areas of the surface of Mars (Kereszturi and Németh, 2012; Robbins et al., 39 2013). On Earth such volcanic rocks often exhibit distinct magnetic anomalies (e.g. 40 Bolos et al., 2012; Urrutia-Fucugauchi et al., 2012). However, only few examples have 41 been mapped with high resolution (e.g. Cassidy and Locke, 2010). 42 Thus, an agglutinated spatter cone of 170 m diameter and surrounding Quaternary 43 lava sheet of the Pali Aike Volcanic Field (PAVF) in southernmost Patagonia (Figs. 1 and 44 3; Skewes and Stern, 1979) has been selected as potential martian analogue. The well-

5

1 preserved morphology and stratigraphy indicates an age of approximately 1.0 Ma 2 when considering the succession of various nearby volcanic formations for which ages 3 of 0.16 to 1.5 Ma have been reported (Mejia et al., 2004). The investigated crater is 4 partly filled by pyroclastic material, eolian sediments as well as blocks and detritus, 5 which have been collapsed from the eastern inner crater wall. 6 The mapping site (52°06’43’’S; 69°42’28’’W; 227 m elevation) covers an area of 7 400 x 400 m, including the crater and its surroundings (Fig. 3A). 8 SITE 3 “Monturaqui”. Impact craters represent a very frequent feature on the Mars 9 surface (Lillis et al., 2013). Depending on e.g. size, target rocks, impactite composition 10 and possible hydrothermal processes they can be characterised by distinct and 11 complex magnetic signatures (e.g. Osinski et al., 2013). On Earth, large impact craters 12 are strongly eroded. In addition, some of them are covered by vegetation or modified 13 by anthropogenic influences. A Late Pleistocene simple type impact crater in the 14 Atacama Desert of northern Chile was selected for this case study (Fig. 1). The crater 15 was discovered in 1962 from aerial photographs and firstly described by Sánchez and 16 Cassidy (1966). It is located at latitude 23°55’40’’S and longitude 68°15’42’’W at an 17 elevation of 2984 m (Ugalde et al., 2007), has a diameter of 370 m and is 34 m deep 18 (Fig. 4 A, B, C), and was formed during the Quaternary (660 ± 90 kyr BP; Ukstins Peate 19 et al., 2010). Due to the arid climate, it remained morphologically well preserved. The 20 crater has remarkable morphological similarities to the Bonneville impact crater on 21 Mars, which was explored by the Spirit rover of NASA (Grant et al., 2004). Monturaqui 22 target rocks include Jurassic granites cut by some mafic dykes. Both rock types are 23 overlain by a several metre thick sheet of Pliocene ignimbrites. Tiny Fe-Ni-Co-P 24 spherules, all bound in impact glass, have been found within the ejecta blanket. They 25 suggest an iron meteorite as impactor (Bunch and Cassidy, 1972; Kloberanz, 2010). 26 SITE 4 “Bahía Glaciares”. Plutonic rocks and layered intrusions form significant parts of 27 the martian crust (e.g. Francis, 2011) and analogues on Earth (McEnroe et al., 2004 28 and 2009). These rocks may have the capacity to store remanent magnetic signatures 29 that can be used to distinguish between different magmatic rock types during future 30 rover-based magnetic surveys. The Patagonian Batholith in the southernmost Andes 31 provides a good example of continental crust formation on Earth and other planets 32 (Behrmann and Kilian, 2003; Diaz-Michelena and Kilian, 2015). A small mapping area 33 of 20 x 6 m (120 m2) was defined on a Cretaceous granite (Fig. 5; Hervé et al., 2007), 34 which is cross-cut by several mafic North-trending (~ 5 °N) more or less parallel mafic 35 dykes (52°48’28’’S; 73°14’10’’W; 11 m a.s.l.). This area was chosen because it is a 36 good example of very low intensity magnetic contrast in a small extension, where 37 transition between alternating mafic and felsic outcrops appears at a centimetre- 38 scale. 39 40 2.4 Additional analyses 41 42 The magnetic field surveys have been complemented with other rock analyses to 43 improve the interpretation of the magnetic signatures of the surveys. Even though the 44 detailed analysis is out of the scope of this work, the types of measurements are 45 briefly described because they support partially some of the conclusions of the work. 46 Firstly, a macroscopic description of the rock types and mineral components has

6

1 been done at each site. Representative rock samples were collected along the tracks 2 for macroscopic investigation and future analyses in the laboratory. For instance, the 3 samples from Pali Aike (Site 2) and Bahía Glaciares (Site 4) have been analysed with 4 polarization and refracted light microscopy of thin sections of the rocks. Texture and 5 grain sizes of samples from el Laco (Site 1) have been also investigated with a scanning 6 electron microscope (Leo 435 VP, Geology Department, Trier University). The mineral 7 composition of granites and amphibolitic dykes of Bahía Glaciares (Site 4) have been 8 analysed by an X-ray diffractometer (Siemens D500, Geology Department, Trier 9 University). 10 Hysteresis properties of representative samples from Pali Aike (Site 2) and Bahía 11 Glaciares (Site 4) have been characterized magnetically at room temperature by means 12 of a vibrating sample magnetometer at the Space Magnetism Laboratory of INTA, 13 Spain. Magnetic susceptibilities have been also measured with a MS-2 susceptometer 14 by Bartington along the transects at Site 4. 15 16 17 3. Results 18 19 The comparative performance and results of the magnetic surveys with MOURA and G- 20 858 magnetometers are described below. In all cases with identical single point 21 measurements the correlation between the two instruments and the two sensors of 22 MOURA has been analysed. The number of differences among the distinct sites has 23 made it possible to demonstrate the versatility of MOURA. Thus, in each study case 24 some of the individual capabilities of the instrument will be highlighted and discussed. 25 26 3.1 SITE 1 “El Laco” 27 The surveys were performed with both instruments using continuous and discrete 28 measurement modes. During the surveys the temperature was ranging from 5°C to 29 27°C. The transects and individual measuring points were georeferenced with the GPS. 30 A Matlab code was used to combine, interpolate and merge the magnetic anomalies 31 and tracks (Fig. 2). The vertical magnetic gradient has been measured at one point 32 where magnetite-bearing ores crop out. Fig. 2B shows an exponential increase of the 33 magnetic anomaly from 1.9 m altitude above the ground down to the rock surface 34 (from + 3,000 to +23,000 nT). The gradient field calculated from the vector 35 components of both vector sensors shows a strong negative vertical component (mz = - 36 11.7 A/m) for this point, which has been modelled by a local shallow superficial dipole 37 with a volume of 0.1 x 0.1 x 0.4 m, an inclination of -62° and a declination of -67°. 38 The magnetite-bearing outcrops exhibit very high positive magnetic anomalies 39 from 30,000 to >110,000 nT while surrounding andesitic lavas and pyroclastic material 40 have much lower positive anomalies (+ 100 to + 2,000 nT). The magnetic anomalies 41 across outcrop transitions between andesites and magnetite-bearing ores have been 42 measured with MOURA in a discrete mode directly on the ground (Fig. 2C) and with a 43 continuous mode (Fig. 2D). The differences in the intensity of andesite anomalies 44 between these two measurements are related to the different and slightly variable 45 distance between the hand-held sensor (25 to 30 cm) and the ground surface of single 46 points (0 cm). The continuous measurements show a large variability of the magnetic 47 anomalies along the ore-bearing outcrops related to either heterogeneous ore

7

1 compositions or slight variations of the sensor distance from the surface. 2 During G-858 surveys the magnetometer became often saturated when high local 3 magnetic anomalies were reached (> 80,000 nT). Local surveys with a higher spatial 4 resolution in a continuous mode showed that this situation appeared very frequently 5 and thus did not permit a complete high-resolution survey with this instrument. The 6 surveys with MOURA were not affected by such saturation since this magnetometer 7 has an extended range mode (auto) that doubles the nominal range to ± 130,000 nT 8 per axis, and thus, allows measurements up to higher field intensities as it has been 9 suggested partly for the martian surface. Despite this problem with the G-858 the 10 scalar magnetic maps of both magnetometers are similar with a correlation R2 > 0.8. 11 (comparison in Figs. 2E and 2F). Of relevant importance is that MOURA magnetometer 12 allows the identification of the component that saturates. 13 Since MOURA magnetometer provides vector magnetic data, it is possible to 14 determine the orientation of the field in the area. This is shown in the rosette of Fig. 15 2G together with the paleodeclinations of other rock samples from El Laco Sur 16 determined by Alva Valdivia et al. (2003). 17 18 3.2. SITE 2 “Pali Aike” 19 A dense grid was performed over the depicted surface with G-858 magnetometer (Fig. 20 3A). In this case, MOURA measurements have been performed with the discrete mode 21 (Fig. 3B) to obtain well-referenced vector data. During the survey, the temperature 22 was oscillating between 7 to 15° C. 23 Fig. 3B compares an interpolated magnetic anomaly map measured with G-858 24 with discrete points from MOURA that are illustrated with a colour code. Both data 25 sets match very well (R2 > 0.82). A 3-D view of the interpolated magnetic anomalies 26 mapped with G-858 is shown in Fig. 3C. It documents the very high positive anomalies 27 of the crater rim (up to + 12,000 nT). A W-E transect of the crater and its surroundings, 28 and its geological features shows two pronounced positive magnetic anomalies 29 centred at both sides of the crater rim where the agglutinated spatter have been 30 mainly deposited as pillow-like blocks of metre size (Fig. 3D). 31 Vector information obtained with MOURA magnetometer at the individual points 32 along the crater rim and crater infill is illustrated in Fig. 3E. In general the predominant 33 declination in all the measurements taken on consolidated lava blocks and the 34 sedimentary infill is around 355°N (white arrows in Fig. 3E). However, anomalous 35 deviations have been detected in the Eastern and Southern part of the crater (red 36 arrows in Fig. 3E), on single basaltic lava blocks, which have removed into the crater 37 during a post-eruptive collapse of the inner crater wall. 38 MOURA has two magnetometers at a small distance of 10 mm between them. 39 They can be used also to measure some of the components of the gradient of the field. 40 Fig. 3G shows the derivatives respect to z of the components Bx, By and Bz of the field. 41 In good agreement with the previous conclusion, the gradient seems to have a 42 homogeneous direction all over the crater with the exception of measurements on 43 single lava blocks that have been removed and re-orientated during the collapse of the 44 wall.

8

1 For, all these data it has been taken into account the tilt angle of MOURA apart 2 from the deviation respect to the North taken with the GPS. This has been possible due 3 to the fact that MOURA has a tilt angle sensor to measure the deviation from the 4 horizontal. This sensor has been used in this example to derive a gravity contrast along 5 the transect within the crater and along its rim which is illustrated in Fig. 3F. Highest 6 values occur along the eastern and western crater rim whereas lowest values are 7 measured at the western eolian sedimentary crater infill. This relationship is shown in 8 the W-E transect of Fig. 3D. 9 10 3.3. SITE 3 “Monturaqui” 11 In this example we performed a dense grid of the centre of the crater as well as the 12 northeastern, eastern and southern part of the crater rim. An interpolated map shows 13 very slight magnetic anomalies (< 50 nT) within the crater (Fig. 4B) whereas more 14 pronounced local negative and positive anomalies from - 400 to > + 600 nT occur 15 within several meters along the crater rim indicating the existence of metre-sized 16 dipoles. The anomalies are not related to outcrops of exposed granitoids and 17 ignimbrites (Fig. 4C). 18 A higher-resolution mapping was performed at a local area of around 10 x 20 m at 19 the north-eastern crater rim with both magnetometers in a continuous mode with G 20 858 and by discrete points with MOURA. Both magnetometers show relatively high 21 positive and negative anomalies ranging from - 3,500 up to > + 6,000 nT (Fig. 4D). This 22 local field of anomalies indicate the existence of not exposed but near-surface metre- 23 sized dipoles. The anomalies measured with both magnetometers are compared in an 24 X-Y plot of Fig. 4E and indicate a correlation of R2 of 0.81. 25 The local mapping area at the north-eastern crater rim is characterised by 26 pronounced local topography changes in the range of ± 5 m of elevation. Fig. 4F 27 illustrates that more pronounced negative anomalies occur at topographic lows. This 28 relationship between lower topography points and higher positive anomalies (and vice 29 versa) was also measured with the B1 and B2 sensors of MOURA and is shown in Fig. 30 4G. This figure documents also the good correlation between both MOURA sensors, B1 31 showing on average + 1,500 to + 2,000 nT higher values than B2 related to the fact that 32 B1 is 10 mm nearer to the ground surface. 33 34 3.4 SITE 4 “Bahía Glaciares” 35 A 20 x 20 m area was mapped with G-858 and MOURA magnetometers along seven 36 high-resolution tracks perpendicular to the dykes (Fig. 5B to 5D) using both continuous 37 and discrete modes. The spacing between the lines is approximately 80 cm and the 38 distance between individual measurement points along the lines range from 5 to 10 39 cm. All the lines show similar patterns, which allow performing an interpolated map of 40 the area: Figs. 5C and 5D show that the dykes clearly contrast with the granites by 41 slight positive anomalies. 42 Fig. 5E shows one of the high-resolution transects where the magnetic signatures 43 have been measured with both magnetometers and a M2 Bartington device for 44 susceptibilities. The dykes exhibit very clear but weak positive anomalies with respect 45 to the granite in the range from + 20 to + 80 nT. Both magnetometers show similar

9

1 patterns (R2 > 0.8). Overall slightly higher values (around 30 nT) of MOURA data are 2 attributed to the fact that they have been performed directly on the rock surface 3 whereas those of G-858 magnetometer are measured at a certain distance to the 4 surface (25 to 30 cm). The susceptibility transect shows a very sharp transition at the 5 interfaces between the granite and dykes, the latter having around 70 x 10-6 SI higher 6 values on average. 7 The magnetic anomalies and the susceptibility data show laterally displaced (Fig. 8 6A, B). This reflects the eastward tilt of the dykes which have dipping angles of 50 to 9 80° and indicates that the uppermost 2-3 m of the mafic dykes are also integrated 10 within the anomalies measured with both magnetometers while the susceptibility 11 shows only the information of the first centimetres below the surface. 12 Detailed petrographic studies including electron microprobe and XRD analysis 13 show that both granites and amphibolitic dykes do not contain magnetite, but instead 14 include ferromagnetic monoclinic C4 pyrrhotite as magnetic carrier. Areal microscopic 15 mapping of pyrrhotite in thin sections of the different dykes and the granite and XRD 16 analyses of the different rocks indicate a content of 1 to 4 vol.% pyrrhotite with grain 17 sizes ranging from < 5 to 150 µm. There is a good correlation between the pyrrhotite 18 contents of the different dykes and the amount of the positive anomalies (Fig. 5E). 19 20 21 4.0 Discussion 22 23 The results of the different ground magnetic surveys of both magnetometers are 24 discussed with respect to the appropriateness of the different instruments and the 25 relationship to mineralogical and magnetic properties of the exposed rocks. In 26 particular the potential of high-resolution detection of weak magnetic contrast 27 between different surface rock types is considered. 28 29 4.1 Instrument Performance 30 Final processed data from both instruments show a very good correlation in intensity 31 of magnetic anomalies (in all cases R2 > 0.8) for the overall measurement range 32 between - 2,000 nT and 100,000 nT (Figs. 2E, 2F, 3C, 4E and 5E). 33 The stability of both instruments has been appropriate for the different 34 surveys. G-858 MagMapper shows a better thermal stability, which can be observed 35 during faster temperature variations during the dawn and dusk, when the transducer 36 may experiment thermal variations up to 0.1 °C/min. The simultaneous measurement 37 of the temperature and the magnetic field during the prospections diminishes this 38 problem, which can be neglected in areas with magnetic anomalies > 100 nT, but the 39 error can be significant (1 %) in low contrast anomalies (1 nT), also due to the 40 resolution of MOURA instrument. Other ways to compensate the temperature effects 41 could improve these errors (Díaz-Michelena et al., 2015 b). For static measurements 42 the simultaneous temperature measurement corrects very well the magnetic field data 43 (Díaz-Michelena et al., 2015 a). 44 Regarding the dynamic range, both magnetometers have also casted 45 appropriate results in most of the cases. The limitation in this feature affects in a 46 different way the response of both instruments. G-858 is influenced in the measured

10

1 modulus of the field, while MOURA is affected separately in every axis. Apart from the 2 extension of the range in modulus, this is an advantage since it could provide useful 3 data in two directions despite of saturation in the other axis. For example, at Site 1, 4 the huge intensity of the anomalies makes it impossible to map them with G-858, 5 while MOURA can measure them in the auto mode, when the maximum offset is 6 applied. 7 At the El Laco site MOURA surveys were performed with discrete and 8 continuous modes. The continuous mode enabled a higher resolution and an easier 9 performance but it may include a shifting by slight variations of the distance between 10 sensor and the ground. The extreme high gradient at this site causes a pronounced 11 fluctuation of around > 5,000 nT when altitude of the sensor changes from 25 to 30 12 cm. This can be avoided by discrete measurements directly on the ground, which also 13 enable a better orientation control of the vector sensor. In the case of highly positive 14 anomalies and in combination with high Q-ratios (remanent versus induced magnetic 15 signatures) of the surface rocks, the vector measurements may also have the capability 16 for paleomagnetic implications which is extremely important for planetary exploration. 17 Surface rock alteration processes which modify the magnetic signatures have 18 influence on limited areas. In particular, the related mineral transformations processes 19 are a direct consequence of the contact of the rocks with the hydrosphere and 20 atmosphere, and their influence depth is limited to several tens of metres. This fact 21 together with the exhumation processes offers often the possibility to correlate the 22 measured direction of the magnetization with the coetaneous paleomagnetic field. In 23 the case of Mars, where the main source of field is the remanent magnetization, 24 oriented measurements does not only contain information on the carriers and the 25 possible alteration effects suffered by them, but also record the paleomagnetic field 26 orientation. This possible tool has been applied to remanence dominated sites which is 27 further discussed in section 4.2. 28 The sensor head orientation is also a matter of discussion. Despite the good 29 signal-to-noise ratio presented by the scalar magnetometer, it is affected by the 30 relative orientation of the head and the magnetic field vector. This is highly improved 31 in a 3-axes magnetometer like MOURA. 32 Another consideration is the gradient immunity and capability to derive a 33 gradient of the field of the instruments. On the one hand, MOURA instrument presents 34 a better gradient immunity, which makes it very suitable to map areas with high 35 frequency patching of the signatures. For example, it is very appropriate to perform 36 decimetre-scale resolution mappings like in the cases of El Laco, Monturaqui and Bahía 37 Glaciares. G-858 presents troubles with not so high gradients (> 20,000 nT). 38 On the other hand, the inclusion of a second head (and therefore to have two 39 3-axes magnetometers) in MOURA instrument offers the capability to better 40 understand the characteristics and depths of the sources. This cannot be applied to 41 deep and extended magnetic sources, because the distance between the two 42 magnetometers is very small (10 mm) but it is useful to analyse near surface 43 heterogeneities and will be discussed in section 4.2. 44 45 46 4.2 Capacity for high resolution mapping with tracing of mineralogical and 47 geological characteristics

11

1 High-resolution ground surveys may indicate compositional variations in soils and/or 2 uppermost crustal rocks, depending on the magnetic contrast between different 3 exposed rocks and the intensity of active magnetic field (Gobashy et al., 2008; Hinze et 4 al., 2013). Despite the fact that it is not the primary goal of MOURA instrument, which 5 is part of the instruments suite of a lander, due to its potential in future exploration 6 missions, the exploration capacity of the commercial G-858 and MOURA 7 magnetometer for extra-terrestrial high-resolution mapping is discussed in the 8 following for the different investigated sites. 9 10 El Laco 11 At this site the intensities in the magnetic anomalies range from 0 to + 110,000 nT 12 (Figs. 2C to 2F) which is unique compared to other magnetic mapping results on Earth 13 (e.g. Hinze et al., 2013). The magnetic contrast at the surface transition between 14 andesitic rocks and magnetite-bearing ores is very sharp and extremely high. A change 15 from + 1,200 to + 80,000 nT appears in less than a metre distance. In general, MOURA 16 data show a better definition of the surface rock transitions and local variabilities in 17 the areas with outcrops of magnetite-bearing ores compared to G-858. Furthermore, 18 G-858 could not register some of the anomalies because its response was occasionally 19 saturated (110,000 nT) (Fig. 2D). Some variations in areas with exposed ores (in the 20 order of ± 5,000 nT) could have been caused by slight changes in the sensor distance 21 from the ground during a continuous measuring mode (see Chapter 3.1), the major 22 variations are probably related to the heterogeneous composition and locally distinct 23 magnetic behaviour of the magnetite-bearing ores. The texture of magnetites in the 24 ores indicates in part a primarily volcanic origin, but also frequent recrystallization 25 during later hydrothermal processes can be observed. It is likely that the hydrothermal 26 crystallization took place over a longer period of the Early Quaternary which may have 27 also included magnetic reversals (e.g. Alva-Valdivia et al., 2003; Naranjo et al., 2010). 28 Our field observations and laboratory analyses of collected samples indicate a large 29 scatter in the grain sizes, the porosity content as well as the relative amount of 30 additional apatite (non-magnetic) and pyroxene in these rocks. These features might 31 explain the observed variations. Other variables reported by Alva-Valdivia et al. (2003) 32 include hysteresis parameters and highly variable Q ratios (from 0.01 to >5,000) 33 indicating a wide range of individual properties of the magnetic carriers, compatible 34 with pseudo-single-domain up to multi-domain status. Therefore, we attribute the 35 observed huge variability in magnetic anomalies to local variabilities in the behaviour 36 and magnetic properties of magnetites in near-surface rocks. 37 In areas where andesitic lavas are exposed, field surveys show only low 38 fluctuations of the positive anomalies (Fig. 2C and 2D). This let us to hypothesize that 39 there are no underlying local ore bodies and the lava flows have relatively 40 homogenous compositions. 41 Measurements with the MOURA vector magnetometer show a clear northward 42 declination between 350° to 10° N (Fig. 2G). This value is similar to the present 43 declination of the IRGF at this site (3° N) and can be explained by a very strong induced 44 magnetization consistent with the very high susceptibilities, but relatively low Q ratios 45 (0.01) measured in six of seven samples from El Laco Sur by Alva-Valdivia et al. (2003). 46

12

1 Pali Aike 2 Magnetic surveys have been performed at some volcanoes world-wide with different 3 spatial resolution, e.g. from Australia (Blaikie et al., 2012), New Zealand (Cassidy and 4 Locke, 2010) and Italy (Okuma et al., 2009). In general these case studies show more 5 and less positive magnetic anomalies (up to a few thousand nT) depending on the 6 composition of the volcanic rocks, its cooling history and the single versus multi 7 domain status of their magnetites (Clark, 1997). Our example of a small crater (170 m 8 diameter) and its surroundings at the Pali Aike Volcanic Field was performed with both 9 magnetometers and with a higher spatial resolution than the previous studies. The 10 transect across the Pali Aike crater shown in Fig. 3C and 3D has a spatial resolution of 11 30-50 cm in-between individual measuring points and thus provides a very good 12 differentiation of different kinds of exposed rocks within the uppermost 1-2 m. 13 Despite the relatively high intensity of the IGRF at Pali Aike (+ 31,000 nT; Fig. 1) the 14 transect across the crater is characterised by very high positive magnetic anomalies of 15 up to + 12,000 nT. These anomalies are mostly pronounced along the crater rim where 16 metre-sized melt spatters have been deposited and cooled down (Fig. 3D). The very 17 strong magnetic signature can be explained by the fact that these lavas contain very 18 frequent tiny magnetite crystals with single domain characteristics in their glassy 19 matrix. The positive anomalies become much lower towards the crater infill and the 20 outer slopes of the crater. An increasing amount of pyroclastic deposits with 21 reorientations during the deposition processes on the steeper slopes of the crater 22 could have reduced the integrated magnetic anomaly of these components. The 23 relatively low local anomalies measured within the crater can be also explained by 24 such multiple re-orientations of magnetic carriers during local redistribution processes 25 including fluvial and eolian activities. 26 Fig. 3E shows arrows for the declination calculated from the vector data of 27 MOURA. The values of all measurements on consolidated lava blocks (white arrows in 28 Fig. 3E) range from 352° to 360° N. These orientations may reflect either the induced 29 present magnetic field or paleofield directions, or a combination of both, depending 30 on their remanence and related Q ratios which reflect often the single versus 31 multidomain status of the basalts. Hysteresis measurements of 25 basaltic lava 32 samples from our mapping area indicate relatively high Q-ratios of 50 to >500 33 suggesting a strong remanence. Comparable basaltic rocks from other sites worldwide 34 are also characterized by a very strong remanence and a predominant single domain 35 status (Day, 1997; Dos Santos et al., 2015; Dunlop, 2002; Zhao et al., 2006). Our 36 measurements which have been performed directly on detached lava and scoria 37 blocks, clearly collapsed from the inner crater wall, show multiple orientations. They 38 are indicated by red arrows in Fig. 3E and include easterly and westerly declinations. 39 The present field which has a declination of 12°N at Pali Aike, the MOURA field 40 vectors as well as several paleodeclinations from different old Pleistocene lavas of Pali 41 Aike (including magnetic reversals; Mejía et al., 2004) have been compiled in Fig. 3H. 42 The estimated age of the investigated cone is approx. 1.0 Ma which suggests a normal 43 global field for that time. MOURA declinations, ranging from 352° to 360° N, are in a 44 very good agreement with such a normal declination as well as other normal 45 declination constrained for other lavas from the Bruhns magnetic period (Mejía et al., 46 2004) rather than the present field declinations which contrast by +15°. This result

13

1 indicates that MOURA magnetometer could provide paleodeclinations when rocks 2 have a very high remanence (high Q ratios). 3 4 Monturaqui impact crater 5 Planetary impact craters can be characterised by a variety of magnetic anomalies 6 which are related in particular to distinct magnetic carriers of the target rocks (e.g. 7 mafic versus felsic or sedimentary) and the sedimentary crater infill as well as the 8 compositions of the impactor, impact-induced melt/glass and/or impact-related 9 hydrothermal mineralization and/or demagnetization (e.g. L’Hereux et al., 2008; 10 Langlais and Thébault, 2011; Osinski and Pierazzo, 2013; Pilkington and Grieve, 1992; 11 Prezzi et al., 2012). 12 At the relatively small Monturaqui crater, a coarse grid of magnetic mapping with 13 spacings of approximately 70 m in-between single measuring points have been 14 previously performed with a caesium magnetometer in the crater and its surroundings 15 by Ugalde et al. (2007). The published interpolated map shows only very low magnetic 16 anomalies in the range of less than ± 200 nT with slightly higher values in the southern 17 and eastern sector of the crater rim. Our grid was measured with a continuous mode 18 of the G-858 and provides a much higher resolution for the central part of the crater as 19 well as its northern, eastern and southern rim and slopes (Fig. 4B). Our data show 20 similar low magnetic anomalies of ± 50 nT for the crater floor as measured by Ugalde 21 et al. (2007), whereas the eastern and north-eastern crater rim is characterised by 22 much stronger local anomalies from -500 to > + 600 nT. The transitions of the outcrops 23 of granitic and ignimbritic target rocks shown Fig. 4C cannot be depicted by the distinct 24 anomalies, while two northwest to southeast-trending dykes seems to responsible for 25 some anomalies at the eastern crater rim (Fig. 4B). In general, the crater rim is 26 characterised by a patchwork of pronounced local positive and negative anomalies 27 which can be caused by small-scale near-surface dipoles which are probably not 28 exposed. Not exposed fragments of the impactor, for which a diameter of ~15 m has 29 been modelled (Echaurren et al., 2005), represent a potential source for these 30 magnetic anomalies. Fe-Ni spherules which have been found in impact melt/glass 31 fragments along the eastern and western crater rim has been classified as a Group I 32 Coarse Octahedrite (Bunch and Cassidy, 1972; Buchwald, 1975) and contain 33 schreibersite, cohenite, fossil taenite, and kamacite as major components as well as 34 some troilite. Laboratory analyses of these components show both high remanent 35 magnetization as well as very high susceptibility (Ugalde et al., 2007; Ukstins Peate et 36 al., 2010). 37 The very pronounced negative to positive anomalies from - 2,000 to + 5,000 nT 38 which have been mapped in a local area of 50 to 100 m extension at the northeastern 39 crater rim (Fig. 4D) also require near-surface rocks with strong dipoles. Not exposed 40 metre-sized fragments of the iron-bearing impactor represent the most likely 41 explanation. In this local area topographic highs are formed by ignimbrite blocks of 42 low-magnetic signature which has been ejected from the crater and deposited on the 43 rim during the impact event. Topographic highs formed by these blocks cause a higher 44 distance of the magnetic sensor with respect to the probably underlying fragments of 45 the impactor while measurements in topographic lows show much higher positive 46 anomalies (Fig. 4F and 4G). 14

1 2 Bahía Glaciares 3 Mafic dykes within felsic to intermediate crustal rocks often produce pronounced local 4 positive magnetic anomalies since they include frequent tiny magnetites (e.g. Hinze et 5 al., 2013). However, in our case study the petrographical investigations indicate that 6 the investigated dykes have not preserved their original magmatic mineral textures 7 and the mineral assemblage point to an emplacement and later equilibration of the 8 dykes under upper greenschist to amphibolite facies conditions (Bucher and Grapes, 9 2011; Philpotts and Ague, 2009). Granites and dykes do not contain magnetite, but 10 both contain monoclinic pyrrothite as magnetic carrier (Dekkers, 1988, 1989; Clark, 11 1984). This mineral appears disseminated and along veins and has been formed during 12 hydrothermal mineralization together with Cu and Au enrichments during the 13 exhumation (Díaz-Michelena and Kilian 2015; Nelson, 1996; Schalamuk et al., 1997). 14 Despite the lower potential of pyrrhotite to produced magnetic anomalies both 15 magnetometers (MOURA and G-858) clearly show high resolution and slightly positive 16 magnetic anomalies (+ 30 to + 80 nT) as well as higher susceptibilities of the dykes 17 compared to the granites (Fig. 5E). The surface transitions between dykes and granites 18 are characterised by very sharp anomalies within a decimetre scale. The amount of 19 pyrrothite (1 to 4 vol. %) which has been quantified in samples of the granite and 20 seven dykes shows a very good correlation with the intensity of the positive anomaly 21 (from + 160 to + 220 nT; Fig. 5F). This result confirms the potential of both 22 magnetometers to explore local mineral enrichments which are often produced by 23 hydrothermal processes associated with gold and copper enrichments (Direen et al., 24 2008). 25 Hinze et al. (2013) show examples of mafic dykes in felsic rocks where different 26 dyke geometries cause distinct shapes of local magnetic anomalies across dykes. For 27 the investigated Bahía Glaciares site Fig. 6 illustrates asymmetric behaviour of the 28 magnetic anomalies along lines which have been measured across dykes which dip 29 with between 50 to 80°. All anomalies are slightly displaced towards the shallower 30 dipping site of the dykes indicating an integration of the magnetic signature of the 31 uppermost 2-3 m of not exposed dykes. This fact together with the difference in 32 contrast obtained with the magnetometers and the susceptometer (Fig. 5) points out 33 that the magnetic field measurements average large volume sources and can lead to 34 wrong conclusions if used as a quantitative mineralogical marker. These results 35 indicate that a multihead (magnetometer and susceptometer) instruments would 36 provide much better results for this purpose. 37 38 39 5. Conclusions 40 41 Several sites with a huge variability in magnetic anomalies have been analysed. As a 42 first conclusion it can be said that the surface measurement of the sourced field often 43 gives direct information on the composition, petrogenesis and alteration processes of 44 the surface rocks.

15

1 For the study, two different magnetometers have been used. On the one hand 2 MOURA vector magnetometer and gradiometer (< 200 g: 72 g instrument + batteries 3 and control PC), developed for Mars surface measurements, as the demonstration of 4 the technology for planetary surveys, and on the other hand a commercial caesium G- 5 858 magnetometer (8 - 9 kg), also used as a reference. The studied areas are 6 considered Mars analogues and they are representative of the intensity range of the 7 expected anomalies on the Red Planet crust. 8 According to the comparison with the reference instrument, it has been 9 demonstrated that MOURA magnetometer is not only appropriate for the static 10 measurement of the absolute value of the magnetic field an its temporal variations, 11 but is also suitable for the prospection measurements in the range of sourced fields 12 from < 15,000 to > 120,000 nT and to reproduce the magnetic contrast of the terrains. 13 Furthermore, MOURA offers vector data of the field and components of the gradient 14 with a significantly lower mass. 15 The particular conclusions for the four case studies are the following: 16 1) El Laco magnetite-bearing ore deposits in the Northern Andes of Chile represents a 17 world-wide unique example with extremely high on ground anomalies ranging from 18 + 30,000 to + 110,000 nT, which may be comparable to highly magnetic rocks of the 19 Noachian martian crust. 20 In this case MOURA enabled better results than G-858 due its larger range of 21 operation (130,000 nT /axis). The declinations measured also by MOURA vectors 22 represent the active global field due to induced-dominated magnetic rock 23 properties with very low Q-ratios. 24 2) A crater in the Pali Aike Volcanic Field, in southern Chile, shows very high positive 25 magnetic anomalies (up to 12,000 nT) of its crater rim caused by frequent tiny and 26 single domain magnetite crystals in the matrix of basaltic lava spatters. 27 Since these rocks, like that of many other comparable volcanic rocks on Earth 28 and other planets, have high Koenigsberger ratios (Q) and thus are dominated by 29 their remanence, MOURA vector data have been used to determine the 30 paleomagnetic orientation during the crater formation around 1 Ma before present. 31 In addition, the different later reorientations of single lava blocks during their 32 collapse from the steep inner crater wall could have been constrained by the vector 33 data. 34 3) The small Monturaqui impact crater in the Atacama Desert of Northern Chile 35 represents an analogue for many other simple type craters, like Bonneville crater on 36 Mars. The granitoid and rhyolitic target rocks have few magnetic carriers and only 37 week magnetic anomalies. Pronounced anomalies along the crater rim indicate 38 metre-sized unexposed remnants of the iron-bearing impactor (octahedrite). 39 Local mapping with a decimetre resolution, with intensities ranging from – 2,000 40 to + 6,000 nT, corroborates the existence of such localised, strong and relatively 41 small size (1 metre) dipoles (iron meteorite fragments) near the surface. 42 4) A site within the Patagonian Batholith of the southernmost Andes provides a 43 window into deeper planetary crustal magnetic signatures. The exposed rocks 44 include granites and mafic dykes that have been partly equilibrated at lower

16

1 amphibolite facies conditions, where all primary magmatic magnetites have been 2 transformed to iron-bearing silicates, and a later hydrothermal mineralization 3 produced pyrrhotite as only magnetic carrier. 4 The freshly exposed transitions between these granites and mafic dykes have 5 been mapped on a decimetre-scale. Despite the very low magnetic contrast from 20 6 to 80 nT both rock types could have been clearly distinguished. In addition, the 7 amount of pyrrhotite, ranging from 1 to 4 vol. %, is well correlated with the positive 8 magnetic anomalies of the dykes. This documents the potential for mapping of 9 hydrothermal mineralization processes as well as associated gold and copper 10 enrichments, even if the magnetic contrast is very low. 11 12 13 Acknowledgements: 14 Authors acknowledge all MOURA MetNet team; in particular, the payloads electronics 15 engineering laboratory for their work with the magnetometer and V. Apéstigue for the 16 technical support. This work was supported by the Spanish National Space Programme 17 (DGI-MEC) through the project AYA2011-29967-C05-01 and the Spanish National Space 18 Program of R&D Externalization through the project PRI-PIBUS-2011-1105. 19 20 References: 21 Acevedo R., Rabassa M., Ponce J., Martínez O., Orgeira M., Prezzi C., Corbella H., 22 González-Guillot M., Rocca M., Subías I. and Vásquez C.: The Bajada del Diablo 23 astrobleme-strewn field, Argentine Patagonia: extending the exploration to 24 surrounding areas, Geomorphology, 169-170, 151-164, 2012. 25 Acuña M. H., Connerney J. E. P., Wasilewski P., Lin R. P., Anderson K. A., Carlson C. W., 26 McFadden J., Curtis D. W., Mitchell D., Reme H., Mazelle C., Sauvaud J. A., 27 d'Uston C., Cros A., Medale J. L., Bauer S. J., Cloutier P., Mayhew M., 28 Winterhalter D., Ness N. F.: Magnetic Field and Plasma Observations at Mars: 29 Initial Results of the Mars Global Surveyor Mission, Science, 279(5357), 1676- 30 1680, 1998. 31 Alva-Valdivia, L.M., Rivas, M.L., Goguitchaichvili, A., Urrutia-Fucugauchi, J., González, J- 32 A., Morales, I, Gómez, S., Henríquez, F., Nyström, J.O. and Naslund, H.R.: Rock- 33 magnetic and oxide microscopic studies of the El Laco iron-ore deposits, 34 Chilean Andes, and implications for magnetic anomaly modelling, International 35 Geology Review, 45 (6), 533 – 547, 2003. 36 Balta, J.B. and McSween, H.Y.: Water and the composition of Martian magmas, 37 Geology, 41(10), 1115 - 1118, 2013. 38 Blaikie, T.N., Ailleres, L., Cas, R.A.F. and Betts, P.G.: Three-dimensional potential field 39 modelling of a multi-vent maar-diatreme — The Lake Coragulac maar, Newer 40 Volcanics Province, south-eastern Australia, Journal of Volcanology and 41 Geothermal Research, 235–236, 70–83, 2012. 42 Bolós, X., Barde-Cabusson, S., Pedrazzi, D., Martí, J., Casas, A., Himi, M. and Lovera, R.: 43 Investigation of the inner structure of La Crosa de Sant Dalmai maar (Catalan 44 Volcanic Zone, Spain), Journal of Volcanology and Geothermal Research, 247–

17

1 248, 37–48, 2012. 2 Breitsprecher, K. and Thorkelson, D.J.: Neogene kinematic history of Nazca–Antarctic– 3 Phoenix slab windows beneath Patagonia and the Antarctic Peninsula, 4 Tectonophysics, 464, 10 – 20, 2009. 5 Breuer, S., Kilian, R., Weinrebe, W., Schörner, D. and Behrmann, J.: Glacier and tectonic 6 control on fjord morphology and sediment deposition in the Magellan region 7 (53°S), Marine Geology, 346, 31 – 46, 2013. 8 Bucher, K., and Grapes, R.: Petrogenesis of metamorphic rocks. Springer, 8th edition, 9 428 pp., Berlin-Heidelberg, DOI 10.1007/978-3-540-74169-5, 2011. 10 Buchwald, W.: Handbook of Iron Meteorites, Their History, Distribution, Composition, 11 and Structure, Volume 3. Berkeley, University of California Press, 1975. 12 Bunch, T.E. and Cassidy, W.A.: Petrographic and electron microprobe study of the 13 Monturaqui Impactite, Contrib. Mineral. Petrol. 36, 95-112, 1972. 14 Carter, J. and Poulet, F.: Ancient plutonic processes on Mars inferred from the 15 detection of possible anorthositic terrains, Nature Geoscience, 6, 1008 – 1012, 16 2013. 17 Cassidy, J. and Locke, C.A.: The Auckland volcanic field, New Zealand: Geophysical 18 evidence for structural and spatio-temporal relationships, Journal of 19 Volcanology and Geothermal Research, 195, 127-137, 2010. 20 Clark, D.A.: Hysteresis properties of sized dispersed monoclinic pyrrhotite grains, 21 Geophysical Research Letters, 11, 173 – 176, 1984. 22 Clark, D.A.: Magnetic petrophysics and magnetic petrography: aids to geological 23 interpretation of magnetic surveys, AGSO Journal of Australian Geology and 24 Geophysics, 17, 83 – 103, 1997. 25 Connerney, J.E.P., Acuña, M.H., Ness, N.F., Kletetschka, G., Mitchell, D.L., Lin, R.L. and 26 Reme, H.: Mars magnetic field - Results from the Mars Global Surveyor, PNAS, 27 102 (42), 14970-14975, 2005. 28 Dare, S.A.S., Barnes, S.-J. and Beaudfoin, G.: Did massive magnetite “lava flows° of El 29 Laco (Chile) form by magmatic or hydrothermal processes? New constraints 30 from magnetite composition by LA-ICP-MS, Mineral Deposita, Doi 31 10.1007/s00126-014-0560-1, 2014. 32 Dekkers, M.J.: Magnetic properties of natural pyrrhotite. Part I: behaviour of initial 33 susceptibility and saturation-magnetization-related rock-magnetic parameters 34 in a grain size dependend framework, Physics of the Earth and Planetary 35 Interiors, 52, 376 – 393, 1988. 36 Dekkers, M.J.: Magnetic properties of natural pyrrhotite. Part II: high and low 37 temperature behaviour of Jrs and TRM as function of grain size, Physics of the 38 Earth and Planetary Interiors 57, 266 – 283, 1989. 39 Díaz-Michelena, M., Sanz, R., Cerdán, M. F. and Fernández, A.B., 2015. Calibration of 40 QM-MOURA three-axis magnetometer and gradiometer, Geosciences 41 Instrumentation, Methods and Data Systems, 4, 1-18, 2015a. 42 Díaz-Michelena, M. and Kilian, R.: MOURA Martian magnetometer potential for high

18

1 resolution magnetic mapping, General Assembly 2013. Geophysical Research 2 Abstracts, Vol. 15, EGU2013-1886, 2013. 3 Díaz Michelena, M., Cobos Arribas, P. and Aroca Hernández-Ros, C. Lock-in amplifiers 4 for AMR sensors, Sensors and Actuators A: Physical 222, 149-159, 2015b. 5 Díaz-Michelena, M. and Kilian, R.: Magnetic signatures of the orogenic crust of the 6 Patagonian Andes with implication for planetary exploration, Physics of the 7 Earth and Planetary Interiors, doi:101016/j.pepi.2015.08.005, 2015. 8 Direen, N.G., Pfeiffer, K.A. and Schmidt, P.W.: Strong remanent magnetization in 9 pyrrhotite: A structurally controlled example from the Paleoproterozoic Tanmi 10 orogenic gold province, northern Australia, Precambrian Research 165, 96 – 11 106, 2008. 12 Dos Santos, E., Gattacceca, J., Rochette, P., Scorzelli, R.B., Fillion, G. et al. Magnetic 13 hysteresis properties and 57Fe Mössbauer spectroscopy of iron and stony-iron 14 meteorites: Implications for mineralogy and thermal history, Physics of the 15 Earth and Planetary Interiors, 242, 50-64, 2015. 16 Dunlop, D.J.: Theory and application of the Day plot (Mrs/Ms versus Hcr/Hc). 2. 17 Application to data for rocks, sediments, and soils, Journal of Geophysical 18 Research, 107 (B3), 10.1029/201JB000487, 2002b. 19 Echaurren, A., Ocampo, C., Rocca, M.C.L.: A mathematic model for the Monturaqui 20 impact crater, Chile, , 68th Annual Meteoritical Society Meeting, 21 5004 p., 2005. 22 Eppelbaum, L.V.: Quantitative Interpretation of Magnetic Anomalies from Bodies 23 Approximated by Thick Bed Models in Complex Environments, Environmental 24 Earth Sciences, 74, (7), 5971-5988, 2015. 25 Eppelbaum L.V. and Mishne A.R: Unmanned airborne magnetic and VLF investigations: 26 effective geophysical methodology of the near future, Positioning, 2(3), 112– 27 133, 2011. 28 Francis, D.: Columbia Hills - An exhumed layered igneous intrusion on Mars? Earth and 29 Planetary Science Letters, 310, 59 – 64, 2011. 30 Gardeweg, M., and Ramírez, C. F.: Hoja Rio , II Region de Antofagasta: 31 Santiago, Chile, Servicio Nacional de Geología y Minería, Carta Geológica de 32 Chile, 66, 89 p., 1985. 33 Gobashy, M.M. and Al-Garni, M.A.: High Resolution Ground Magnetic Survey (HRGM) 34 for determining the optimum location of subsurface dam in Wadi Nu’man, 35 Makkah Al Mukarammah, KSA, JKAU: Earth Sci., 19, 57-83, 2008. 36 Grant, A. J., Arvidson, R., Bell III, J.F., Cabrol, N.A., Carr, M.H., Christensen, P., 37 Crumpler, L., Des Marais, D. J., Ehlmann, B. L., Farmer, J., Golombek, M., Grant, 38 F.D., Greeley, R., Herkenhoff, K., Li, R., McSween, H. Y., Ming, D. W., Moersch, 39 J., Rice Jr., J. W., Ruff, S., Richter, L., Squyres, S., Sullivan, R. and Weitz, C. : 40 Surficial Deposits at Gusev Crater Along Spirit Rover Traverses, Science, 305 41 (5685), 807–810, 2004. 42 Greenough, J.D. and Yacoby, A.: A comparative geochemical study of Mars and Earth 43 basalt petrogenesis, Can. J. Earth Sci., 50, 78-93, 2013.

19

1 Groschner, C., Lan, S., Wise, A., Leary, A., Lucas, M., Park, C., Laughlin, D., Díaz- 2 Michelena, M. and McHenry, M.: The role of atmosphere on exsolution 3 processes and magnetic properties of ulvospinel, IEEE Transactions on 4 Magnetics 49(7), 4273 – 4276, 2013. 5 Hervé, F., Pankhurst, R.J., Fanning, C.M., Calderón, M. and Yaxley, G.M.: The South 6 Patagonian batholith: 150 my of granite magmatism on a plate margin, Lithos, 7 97, 373 – 394, 2007. 8 Hinze, W.J., Von Frese R.R.B. and Saad, A.H.: Gravity and magnetic Exploration, 9 Principles, Practices and Applications, Cambridge University Press, 525 p., 2013. 10 Ialongo, S., Fedi, M. and Florio, G.: Invariant models in the inversion of gravity and 11 magnetic fields and their derivatives. Applied Geophysics, 110, 51–62, 2014. 12 Jonnsson, E., Troll, V.R., Högdahl, K., Harris, C., Weis, F., Nilsson, K. and Skelton, A. 13 Magmatic origin of giant ‘Kiruna-type’ apatite-iron-oxide ores in Central 14 Sweden, Scientific Reports, 3, 1644, Doi:10.1038/srep01644, 2013. 15 Kereszturi, G. and Németh, K.: Monogenetic Basaltic Volcanoes: Genetic Classification, 16 Growth, Geomorphology and Degradation, In: Németh, K. (ed.): Updates in 17 Volcanology - New Advances in Understanding Volcanic Systems, ISBN: 978- 18 953-51-0915-0, InTech: 3-88. DOI: 10.5772/51387, 2012. 19 Kilian, R. and Behrmann, J. H.: Geochemical constraints on the sources of continent- 20 related deep sea sediments and their recycling in arc magmas of the Southern 21 Andes, Journal of Geol. Soc. London, 160, 57-70, 2003. 22 Kloberdanz, C. M.: Geochemical analysis of the Monturaqui Impact Crater, Chile, Ph 23 thesis, University of Iowa, 2010. 24 Lan, S., Groschner, C., Runco, J., Wise, A., Díaz-Michelena, M., Laughlin, D. and 25 McHenry, M.: Phase identification and temperature-dependent magnetization 26 of Ti-rich titanomagnetite in different atmospheres magnetics, IEEE 27 Transactions on Magnetics, 49 (7), 4314 – 4318, 2013. 28 Langlais B. and E. Thébault, E.: Predicted magnetic signatures of Martian 29 (de)magnetized impact craters and a statistical comparison with observations, 30 Icarus, 212(2), 568-578, 2011. 31 L’Heureux, E., Ugalde, H., Milkereit, B., Boyce, J., Morris, W., Eyles, N., and Artemieva, 32 N.: Using vertical dikes as a new approach to constraining the size of buried 33 craters: An example from Lake Wanapitei, Canada, in Kenkmann, T., Hörz, F., 34 and Deutsch, A., eds., Large meteorite impacts III: Geological Society of America 35 Special Paper 384, p. 43 – 50, 2005. 36 Lillis, R.J., Robbins, S., Manga, M., Halekas, J.S. and Frey, H.V.: Time history of the 37 Martian dynamo from crater magnetic field analysis, Journal Geophysical 38 Research: Planets, 118, 1–24, doi:10.1002/jgre.20105, 2013. 39 Maksaev, V., Gardeweg, M., Ramírez, C. F. and Zentilli, M.: Aplicación del método 40 trazas de fisión (fission track) a la datación de cuerpos de magnetita de El Laco 41 e Incahuasi en el altiplano de la región de Antofagasta, Congreso Geológico 42 Chileno 5th, Santiago, v. 1, p. B1–B23, 1988.

20

1 McEnroe, S.A., Brown, L.L. and Robinson, P. (2004): Earth analog for Martian magnetic 2 anomalies: remanence properties of hemo-ilmenite norites in the Bjerkreim- 3 Sokndalintrusion, Rogaland, Norway, Journal of Applied Geophysics 56: 195– 4 212. 5 McEnroe S.A., Fabian, K., Robinson, P., Gaina, C. and Brown, L.L.: Crustal magnetism, 6 lamellar magnetism and rocks that remember, Elements, 5, 241 – 246, 2009. 7 Mejia, V., Opdyke, N.D., Vilas, J.F., Singer, B.S. and Stoner, J.S.: Plio-Pleistocene time- 8 averaged field in southern Patagonia recorded in lava flows, Geochemistry 9 Geophysics Geosystems, 5(3), DOI:10.1029/2003GC000633a, 2004. 10 Morschhauser, A., Lesur, V. and Grott, M. A spherical harmonic model of the 11 lithospheric magnetic field of Mars, Journal of Geophysical Research, 119, 6, 12 1162-1188, 2014. 13 Naranjo, J.A., Henríquez, F. and Nyström J.O.: Subvolcanic contact metasomatism at El 14 Laco Volcanic Complex, Central Andes, Andean Geology, 37, 110-120, 2010. 15 Nelson, E.P.: Suprasubduction Mn mineralization: Metallo-tectonic Terranes of the 16 Southernmost Andes, pp. 315-330, In Geophysical Monograph 98, Eds.: Bebout, 17 G.E. et al., 1996. 18 Okuma, S., Stotter, C., Supper, R., Nakatsuka, T., Furukawa, R. and Motschka, K.: 19 Aeromagnetic constraints on the subsurface structure of Stromboli Volcano, 20 Aeolian Islands, Italy, Tectonophysics, 478, 19–33, 2009. 21 Osinski, G.R. and Pierazzo, E.: Impact cratering: Processes and products, Wiley, 330 p., 22 2013. 23 Philpotts, A. and Ague, J.: Principles of Igneous and Metamorphic Petrology, 684 p., 24 Cambridge University Press, 2009. 25 Pilkington, M. and Grieve, R.: The geophysical signature of terrestrial impact craters, 26 Reviews of Geophysics, 30, 161-181, 1992. 27 Robbins, S.J., Hynek, B.M., Lillis, R.J. and Bottke, W.F.: Large impact crater histories of 28 Mars: The effect of different model crater age techniques, Icarus, 225, 173– 29 184, 2013. 30 Prezzi C., Orgeira M., Acevedo R., Ponce J., Martinez O., Rabassa J., Corbella H., 31 Vásquez C., González-Guillot M. and Subías I.: Geophysical characterization of 32 two circular structures of Bajada del Diablo (): indication of impact 33 origin, Physics of the Earth and Planetary Interiors, 192-193, 21-34, 2012. 34 Sánchez, I. and Cassidy, W.: A previously undescribed meteorite crater in Chile, Journal 35 of Geophysical Research, 71, 4891-4895, 1966. 36 Schalamuk, I.B., Zubia, M., Genini, A. and Fernández, R.R.: Jurassic epithermal Au-Ag 37 deposits of Patagonia, Argentina, Ore Geology Reviews 12, 173 – 186, 1997. 38 Sillitoe, R.H. and Burrows, D.R.: New field evidence bearing on the origin of the El Laco 39 magnetite deposits, Northern Chile, Economic Geology, 97, 1101-1109, 2002, 40 Skewes, M.A. and Stern C.R.: Petrology and geochemistry of alkali basalts and 41 ultramafic inclusions from the Pali-Aike Volcanic Field in southern Chile and the

21

1 origin of the Patagonian Plateau lavas, Journal of Volcanology and Geothermal 2 Research, 6, 3-25, 1979. 3 Stolpe, E.M., Baker, M. B., Newcombe, M. E., Schmidt, M. E., Treiman, A. H., Cousin, A., 4 Dyar, M. D., Fisk, M. R., Gellert, R., King, P. L., Leshin, L., Maurice, S., McLennan, 5 S. M., Minitti, M. E., Perrett, G., Rowland, S., Sautter, V., Wiens, R. C. , MSL 6 Science Team: The petrochemistry of Jake_M: A Martian Mugearite, Science, 7 341, 6153. DOI: 10.1126/science.1239463, 2013. 8 Ugalde, H., Valenzuela, M. and Milkereit, B.: An integrated geophysical and geological 9 study of the Monturaqui impact crater, Chile, Meteoritics & Planetary Science, 10 42(12), 2153-2163, 2007. 11 Ukstins Peate, I., Kloberdanz, C., Peate, D.W., Chung Wan, L., Cabrol, N., Grin, E., 12 Piatek, J. and Chong, G.: Non-modal melting of target rocks to produce >ite at 13 Monturaqui Crater, Chile, 41st Lunar and Planetary Science Conference, 1-5 14 March 2010, Houston, Texas, 2010. 15 Urrutia-Fucugauchi, J., Trigo-Huesca, A. and Pérez-Cruz L.: Magnetic links among lava 16 flows, tuffs and the underground plumbing system in a monogenetic volcano, 17 derived from magnetics and paleomagnetic studies, Physics of the Earth and 18 Planetary Interior, 212-213, 10-18, 2012. 19 Valesco, F. and Tornos, F.: Insights on the effect of hydrothermal alterartion in the El 20 Laco magnetitre deposits, Revista de la Sociedad Espanola de Mineralogía, 16, 21 2012. 22 Wray, J.J., Hansen, S.T., Dufek, J., Swayze, G.A., Murchie, S.L., Seelos, F.P., Skok, J.R., 23 Irwin III, R.P. and Ghiorso, M.S.: Prolonged magmatic activity on Mars inferred 24 from the detection of felsic rocks, Nature Geoscience, 6, 1013 - 1017, 2013. 25 Zhao, Xixi, et al. Unraveling the magnetic carriers of igneous cores from the Atlantic, 26 Pacofoc, and the southern Indian oceans with rock magnetic characterization, 27 Physics of the Earth and Planetary Interiors, 156, 294-328. 28 29 30 31

22